奇创岛

From «Nonviolent Communication: A Language of Life»

Build a Personal NVC Communication AI Coaching Toolkit

You'll create one ready-to-use AI prompt template for each of 3–5 high-frequency conflict scenarios, letting AI "translate" what you say in the heat of the moment into NVC expressions — or guide you through post-conflict reflection step by step — forming a personal NVC toolkit you can call on anytime.

Final work

A complete 'NVC Communication AI Coaching Toolkit'

Estimated time

1–2 hr

Submitted

…

Your final work

Purpose:Let AI serve as your NVC translator and reflection coach at critical communication moments — converting emotionally charged language into language that fosters connection, reducing the toll that everyday conflict takes on relationships.

Parts:

  • Toolkit usage guide (which scenarios it fits, how to use it, and its boundaries)
  • 3–5 complete AI prompt templates for real conflict scenarios (each including a system prompt + user-input format + output-format requirements)
  • At least 1 full usage example (demonstrating the complete AI output after entering a real emotionally charged statement)
  • Toolkit reflection (which template works best, what still needs adjustment, and long-term usage notes)

Use cases:

  • · Reflect independently after a family argument, without needing the other person's cooperation
  • · Draft your wording before a workplace conflict and let AI check whether it contains judgmental language first
  • · Run an important message to your partner or child through the rewriting tool before you send it
  • · Use AI to help you find a gentle yet clear way to say no
  • · Build a long-term NVC habit by using AI as a low-cost daily practice partner

Pick a topic

Pick the topic closest to you, or write a custom one when you submit.

Personal Life

Family / Parenting

Work / Projects

Relationships & Communication

Tools you'll use from the book

NVC Four-Component System Prompt

Write the four NVC components (Observation / Feeling / Need / Request) into AI's system role instruction so AI responds using this framework throughout the entire conversation.

How to use it here:

When building each toolkit prompt, your system prompt must include: 'You are a Nonviolent Communication (NVC) coach. Respond using only the four NVC components — Observation (judgment-free facts), Feeling (genuine emotion words), Need (universal human needs), and Request (a specific, refusable action) — and label which component each part of your response corresponds to.' This is the foundational layer for every tool.

Boundaries:

The more specific the system prompt, the better the results — but don't instruct AI to 'lie for you' or to 'make the other person feel they have no choice.' Those requests go beyond NVC and can easily turn the tool into a manipulation device.

Feeling / Need Extraction Prompt

An input template specifically designed to extract genuine feeling words and underlying needs from emotionally charged statements, letting AI do the 'emotional decoding' for you.

How to use it here:

Sample format: 'I just said — or almost said — to [person]: [original statement]. Please help me extract: 1) the genuine feelings I likely had (use emotion words only, not "I feel like you..."); 2) the needs I most likely had behind those feelings (drawing from universal needs such as safety, connection, respect, autonomy, efficiency, etc.).'

Boundaries:

The feelings and needs AI extracts are only candidates — you need to confirm which ones are genuinely true for you. Don't take AI's output of 'you may need...' as fact; instead, ask yourself: 'Does that actually ring true?'

Expression Rewrite Prompt

The core template for rewriting jackal language (judgment / blame / commands) into giraffe language that the other person can actually hear.

How to use it here:

Sample format: 'Please rewrite the following expression in NVC language. Original statement: [what you want to rewrite]. Context: [who you're speaking to and why you said it]. Requirements: 1) the rewrite must include Observation + Feeling + Need + Request; 2) the Request must be specific and something the other person can decline; 3) output the original and rewritten statements side by side, with each part labeled by its NVC component.'

Boundaries:

The rewritten expression should still clearly convey your core message — don't let 'NVC-ifying' it make the meaning vague or evasive. If the rewritten version feels like 'that's not what I was trying to say,' keep adjusting.

Conversation Reflection Prompt

After a conflict, paste in both sides of the conversation and let AI analyze sentence by sentence at what point the exchange shifted from connection to judgment.

How to use it here:

Sample format: 'Please analyze the communication patterns in the following conversation (me vs. [other person]): [paste conversation]. Please identify: 1) which statements are jackal language (containing judgment / comparison / blame-shifting / commands); 2) what feelings and needs might lie behind those statements; 3) which turning point in the conversation was where the connection broke down; 4) if I could do it over, which statement could I replace with an NVC expression?'

Boundaries:

The reflection tool is only for analyzing your own expressions — don't use AI to 'judge who was right or wrong.' That reinforces the jackal pattern. Focus on: 'What different choice could I have made in that moment?'

Self-Empathy Guidance Prompt

When your emotions are running high but you don't know what you need, this structured-questioning tool guides you through an internal NVC dialogue on your own.

How to use it here:

Sample format: 'I'm feeling [upset / hurt / angry / distressed] about [person / situation]. Please guide me through the NVC self-empathy steps: 1) first help me find one emotion word that accurately describes how I feel right now; 2) ask me three questions to help me find the need behind this feeling; 3) once I've identified my need, ask me: "What is one specific request I most want to make of this person right now?" — please go one step at a time, and wait for my response before moving to the next step.'

Boundaries:

The self-empathy tool is only for solo use after you've calmed down. If your emotions are still intense, take a physical pause first (leave the room, drink some water, take a few deep breaths), then open this tool.

Work rules

Your work MUST include

  • At least 3 complete AI prompt templates for real conflict scenarios (not single-sentence questions, but full templates including a system prompt + user-input format + output-format requirements)
  • At least 1 complete usage example (showing a real emotionally charged statement entered into the tool and the full AI output)
  • A toolkit usage guide (which scenarios it fits, when it is *not* appropriate, and its usage boundaries)
  • Each tool template must reflect at least 2 of the 4 NVC components (Observation / Feeling / Need / Request)
  • A toolkit reflection (which template works best, what still needs refinement)
  • The toolkit must target your own real high-frequency situations, not a pile of generic templates

Your work CANNOT just be

  • Don't just explain the four NVC components without providing prompt templates that can be copy-pasted and used directly
  • Don't let AI 'manipulate' the other person, 'make them unable to refuse,' or 'make them think it was their own idea'
  • Don't build a toolkit that uses AI to generate your entire work for you — AI is your translator and coach, not your ghostwriter
  • Don't omit a usage-boundary statement (which scenarios require real face-to-face conversation, not just AI)
  • Don't submit only a generic list of tools with no personalization for your own real situations

AI can help you here

Round 1: Help me choose the scenarios most worth building a tool for

When to use: You're not sure which communication scenarios to build tools for, or you have too many scenarios and don't know where to start.

I'm working on the '{{route name}}' project using *{{book title}}*, with the goal of building a personal NVC Communication AI Coaching Toolkit.

Based on my situation, please help me identify the 3 communication scenarios most worth building tools for first, and explain which type of tool fits each scenario (reflection tool / translation tool / rehearsal tool / self-empathy tool).

My situation:
[Describe your highest-frequency communication struggles, the relationship scenarios most likely to trigger your emotions, and the types of situations you usually regret after a conflict]

Please output:
1. The 3 scenarios I'd most recommend building tools for first (with reasons)
2. The most suitable tool type for each scenario
3. Priority recommendation: which to build first, and which to tackle after you have some experience
4. What will change in your day-to-day communication once you're using these 3 tools well

Yellow placeholders need you to fill in before using the AI.

AI can help you organize ideas, but cannot make final judgments for you. Don't let AI fabricate experiences, cases, or misleading content.

Round 2: Help me build a specific NVC prompt template

When to use: You've already chosen a scenario but aren't sure how to translate the four NVC components into a system prompt and input format that AI can work with.

My project is the '{{route name}}' route in *{{book title}}*.

I want to build an AI prompt tool for the following scenario:
{{topic}}

Please help me design a complete prompt template for this tool with the following requirements:
1. System prompt: clearly tell AI its role, to use only the four NVC components (Observation / Feeling / Need / Request), not to judge the other person on my behalf, and to label which component each part of the output corresponds to
2. User-input format: tell the user what information to provide (original statement / person involved / scene context / desired outcome)
3. Output-format requirements: which sections AI should produce (e.g., component identification + rewritten version + usage tip)
4. Boundary note: which scenarios this tool is *not* suitable for

Please output:
- The complete system prompt (ready to copy-paste)
- The user-input template (with placeholder descriptions)
- A sample expected-output format (demonstrated with a fictional example)
- Tool description (2–3 sentences)

Yellow placeholders need you to fill in before using the AI.

AI can help you organize ideas, but cannot make final judgments for you. Don't let AI fabricate experiences, cases, or misleading content.

Round 3: Help me review the complete toolkit

When to use: You've finished a draft of the toolkit and are ready to submit. You want to confirm the toolkit is genuinely usable, not just NVC knowledge repackaged.

I'm submitting my project work for the Shufang Island project.

Book title: *{{book title}}*
Project route: {{route name}}
My topic: {{topic}}

My toolkit draft:
{{work draft}}

Please check against each of the following criteria:
1. Does the toolkit include at least 3 complete prompt templates (with system prompt + user-input format + output-format requirements)?
2. Does each system prompt include a clear instruction to use the four NVC components — not just a vague phrase like 'please use nonviolent communication'?
3. Is the user-input format clear enough that users know exactly what information to provide?
4. Is there at least 1 complete usage example (real input + actual AI output)?
5. Does the toolkit include a usage-boundary statement (when it is *not* appropriate to use an AI tool)?
6. Does it clarify which parts AI cannot replace (the parts the user must complete in real conversation)?
7. Is the toolkit tailored to the user's real situations, rather than being a pile of generic templates?
8. Is it ready to submit?

Please output:
- Overall assessment (one sentence)
- Item-by-item check results (pass / needs revision + reason)
- Things that must be changed
- Things that could be strengthened (e.g., adding a tool for a high-value scenario)
- Revision suggestions

Yellow placeholders need you to fill in before using the AI.

AI can help you organize ideas, but cannot make final judgments for you. Don't let AI fabricate experiences, cases, or misleading content.

Submit my work